search

guicho271828 / trivia

Pattern Matcher Compatible with Optima
Other
334 stars 22 forks source link

Trivia's atrophied on ABCL #105

Closed remexre closed 5 years ago

remexre commented 5 years ago

Check the latest Travis results (e.g. #438.5) to see an UNBOUND-VARIABLE on load.

It might be a better solution to disable the CFFI tests for ABCL with #-abcl than to allow failures in general.

remexre commented 5 years ago

It looks like this is because trivial-cltl2 assumes ABCL exports all the relevant symbols from the lisp package, when in fact it only exports compiler-let.

guicho271828 commented 5 years ago

hmm, I am not particularly fond of hiding test failures. I think allowing failures is a better option, assuming that ABCL will eventually implement other cltl2 functions.

remexre commented 5 years ago

I mean, right now the tests are failing, just silently; I'm just suggesting limiting the scope of what's ignored.

I'm not sure what the timeline on ABCL getting these functions looks like; https://github.com/armedbear/abcl/issues/87 exists but is closed.

I'm not familiar enough with this codebase to know, would it be practical to change #-(or ecl) to #-(or abcl ecl) here and here? I tried it in my fork, and it looks like it works (at the level of testing I could do during lunch); I can PR it if you want.

guicho271828 commented 5 years ago

change #-(or ecl) to #-(or abcl ecl) here and here?

oh that makes sense. Sure, send me a PR.

remexre commented 5 years ago

Just opened https://github.com/guicho271828/trivia/pull/106; are you open to reworking testing so a regression here fails the Travis build?

guicho271828 commented 5 years ago

CFFI seems to work on abcl after disabling the cltl2 part.